Article

Theme: Author Meets Critics

Free Will Skepticism and Obligation Skepticism: Comments on Derk Pereboom’s Free Will Agency, and Meaning in Life

Dana Kay Nelkin

Email: [email protected]

I. Introduction

Derk Pereboom’s Living Without Free Will (2001) is an outstanding book that rightly changed the debate over an ancient problem. I am certain that Free Will Skepticism, Agency, and Meaning in Life will join the first book as a must-read for anyone in the world interested in these foundational issues. The main pillars of the account remain, including incompatibilism between free will and determinism, skepticism about free will on a combination of conceptual and empirical grounds, and the thesis that life on the assumption that we do not have free will can be meaningful and perhaps even better in certain important respects than life on the assumption that we have it. At the same time, there are some subtle shifts in particular commitments, new ways of framing certain issues, and detailed responses to a good number of the very many challenges to the view as originally set out in Living Without Free Will and related articles. The responses to challenges serve to elaborate, extend, and deepen the view. The result is a book that, along with the first, will be a centerpiece of the conversation about the metaphysics of free will, about morality, about moral responsibility and our related emotions and practices, about the pressing practical and moral issues concerning punishment, and about meaning in life. Finally, the book itself is a wonderful manifestation of a commitment to certain intellectual values--including seeing philosophy as an inclusive conversation, and to testing and revising one’s views in the search for truth.

As can be seen from Gregg Caruso’s précis, the book contains a number of arguments for commitments that fit together. Some are modular, and one could accept one argument without accepting others. At the same time, many are mutually supporting, sometimes in surprising ways that Pereboom skillfully brings out. In the remainder of this paper, I will explore one thread that connects a number of distinct arguments. The particular connection is not always discussed explicitly in the book, and I hope that by bringing it out I can invite Pereboom to elaborate further on the details of his view. In section II, I briefly review key parts of the view that I think are connected in an interesting way, and in sections III, IV, and V raise some questions for the view that emerge in thinking about the parts together.

II. Blame, Punishment, Relationships, and Obligation

To begin, note that Pereboom’s rejection of free will is a rejection of a particular kind of freedom. It is the “strongest sort of control in action required for a core sense of moral responsibility” (p. 2). And the sense of moral responsibility at issue is to be understood in terms of “basic desert”. As Pereboom explains:

For an agent to be morally responsible for an action in this sense is for it to be hers in such a way that she would deserve to be blamed if she understood that it was morally wrong, and she would deserve to be praised if she understood that it was morally exemplary. The desert at issue here is basic in the sense that the agent would deserve to be blamed or praised just because she has performed the action, given an understanding of its moral status, and not, for example, merely by virtue of consequentialist or contractualist considerations (p. 2).

Somewhat unusually for an incompatibilist, Pereboom accepts that Frankfurt cases challenge the Principle of Alternate Possibilities for blameworthiness (PAP-B), the principle that one is blameworthy—in the basic desert sense—for what one does only if one could have done otherwise (see chapter 1 and p. 142, note 12).1 Thus, it is not because it deprives us of alternatives that determinism undermines free will. Rather, it undermines free will because it undermines the idea that we are the source of our actions in the right way, which entails our being agent causes. Pereboom suggests that while our being such agent causes is “prima facie positively conceivable,” to use Chalmers’ terminology, we do not know whether it is “ideally negatively or positively conceivable (p. 58).” But either way, the empirical evidence is such that whether determinism or indeterminism is true, we are not sources of our actions in the relevant sense. Thus, skepticism about blameworthiness in the basic desert sense follows in the end.

While this skeptical conclusion can seem to threaten our entire practices of blaming and praising, the justification for responding in effective ways to criminal behavior, and the assumptions on which our most valuable personal relationships are based, Pereboom offers arguments in each case that these threats are merely apparent. In the case of our practices of blaming and praising, Pereboom argues that while aspects of the practices—notably, those that depend on retributive assumptions—are not justified, other aspects, captured in a well worked out forward-looking account of blame, remain untouched by free will skepticism (chapter 6). When it comes to the treatment of criminal behavior, Pereboom argues that while we cannot justify punishment, we can nevertheless justify incapacitation of criminals on the grounds of the rights to self-defense of other members of society (chapter 7). And when it comes to personal relationships, we can preserve these, as well, even if we disavow or eliminate the attitudes that presuppose free will in the sense required for basic desert (chapter 8).

Each of the three arguments taken individually is powerful, and I am persuaded that something like Pereboom’s account of what justifies, and of what constitutes justified treatment of criminal behavior is correct, as is something like the account of what personal relationships require (or, rather, do not require) in the way of moral blameworthiness and retributive emotions. That is, I am inclined to accept Pereboom’s arguments that the incapacitation of criminals is best justified on non-retributive grounds, and that we can—at least in principle—have the most valuable kinds of personal relationships without a susceptibility to retributive emotions. At the same time, his explication of the forward-looking account of blame reveals that he brings to these arguments an assumption that I am not (yet, anyway) willing to make. To see this, we need to look a little more closely at the forward-looking account of blame.

The account parallels Michael McKenna’s conversational model in that it recognizes three aims for blame (and praise): moral formation, protection, and reconciliation.2 For example, we can blame insofar as we ask agents to explain their reasoning, express our view that what they did was wrong, and request that agents strive to eliminate dispositions to do wrong, thereby serving these aims. This sort of blame captures much of what we actually do when we blame, without presupposing that the object of blame acts freely in the sense required for basic desert. The details of Pereboom’s account are worthy of much more discussion, but I will here turn directly to an objection that Pereboom considers:

Against this skeptical account one might object that for an agent to be subject to blame for performing action A even in this forward-looking sense requires that it was the case that she ought not to do A, and this in turn requires, by the ‘ought’ implies ‘can’ principle (OIC) that she could then have refrained from doing A instead, which is incompatible with determinism. While my free will skepticism does not endorse determinism, it leaves open determinism as a serious possibility. And according to the consequence argument determinism is incompatible with being able to do otherwise, at least in one salient sense, and on this understanding I find the argument attractive (van Inwagen 1983; Ginet 1990: 95-117). So the objection poses a threat to my proposal. (p. 138)

The problem arises because it seems that even blame in the forward-looking sense described by Pereboom requires that it be the case that the agent did something she ought not have done. And if we also accept, as Pereboom does here, the ‘ought’ implies ‘can’ principle (OIC), then having done something one ought not have done requires that one could have refrained from doing it. But if determinism is true—and it is a serious possibility on Pereboom’s view—then one can’t have done otherwise than one actually does. Thus, if determinism is true, then we can never truly say of someone that they did something that they ought to have refrained from doing. And given that determinism is a serious possibility, it is hard to see that we are justified in blaming anyone, even in the forward-looking way that Pereboom articulates.

Now it is worth pausing for a moment to note the significance of the commitment on Pereboom’s part to OIC and to the claim that determinism, if true, would undermine claims of moral ‘ought’ (at least in the sense operative in OIC). As Pereboom notes, those who accept Frankfurt cases as challenging PAP-blame have a natural way of challenging OIC. John Martin Fischer (2003) argues in this way, for example. In the Frankfurt case in which Jones kills Smith, if Jones is responsible for killing Smith, despite Black waiting in the wings, Jones also does something he ought not do. But if he cannot refrain from killing, and OIC is true, then it cannot be that Jones ought to have refrained from killing Smith. Thus, OIC must be false. I am inclined to go in the opposite direction, and I argue that Frankfurt cases do not undermine either PAP-blame or OIC when we use a univocal and relevant notion of “can” or “ability”. Pereboom disagrees with both of these approaches, however, arguing that while Jones is blameworthy in a case in which he cannot do otherwise, it is not the case that he did something that he ought not have done. In other words, he rejects PAP-blame, while accepting OIC.3 Pereboom offers a consistent and intriguing rationale: when it comes to “ought” claims of the relevant kind, they are action-guiding, and simply do not apply when one cannot do otherwise. But judgments of blameworthiness are not similarly action-guiding, and so it could be correct to blame someone in a case in which he could not do otherwise as long as he acted freely in the sense required for basic desert.

Taking this middle path leaves Pereboom to face the original objection head-on. How can we be justified in blaming people--even in the forward-looking sense--if we can’t even tell them they’ve done something they ought not have done? His response draws on several accounts of different senses of “ought.”4 While acknowledging that if determinism is true, we cannot truly say that people have not acted as they ought in an action-guiding sense of “ought” (or in a sense of “specific action demand”), there are other senses of “ought” that are still available to us. In particular, we can use the “ought” of axiological evaluation. We can say that things “ought to be different” than they are, that it would be better if they were different. We can also recommend that an agent act differently in the future by using the “ought” of “axiological recommendation.” Thus, the proposal for a forward-looking practice of blame becomes somewhat more revisionary, but, as Pereboom argues, can still fulfill the three specified aims of our blaming and praising practices. Pereboom also resists the idea that we must jettison talk of “wrongness” if we reject the application of “ought” in the specific action demand sense. Thus, on his view, there is much we can retain, including the practices which contribute toward the three aims of our responsibility practices, as well as the use of “wrong” and “ought”, where the latter is understood in the sense of axiological recommendation.

It is at this point that I would like to raise two sets of questions. The first is whether the original objection to the forward-looking account of blame is, in a way, larger than it seems at first. In particular, the objection might not be conditional simply on the truth of determinism, but instead might remain for Pereboom’s forward-looking account of blame, even if determinism is false. The second is whether Pereboom’s answer to the objection deprives us of some important resources that otherwise help support the three arguments that collectively show that free will skepticism is not so troubling after all, and in fact might show that we are better off than we thought. In the next section, I take up the first set of questions, and in sections IV and V, I take up the second set.

III. What Kind of “Can” in Ought Implies Can?

As Pereboom poses the problem for the forward-looking account of blame, the problem is conditional on determinism. If OIC is true and if it is true that the forward-looking account depends on the claim that agents are blameworthy only when they do what they ought not have done, then the account cannot have application in a deterministic world. This implication is bad enough for Pereboom to reject the claim that the forward-looking account does depend on the ought-claim. But is the problem is limited to deterministic worlds?

Recall from Caruso’s précis that in chapters 2 and 3, Pereboom argues that event causal views cannot solve the disappearing agent problem. Even if the occurrence of a particular decision is about 50 percent probable, on the event causal account, only events are causally relevant. And so nothing about the agent settles whether the decision occurs (p. 50). In other words, if we replace determinism with indeterminism in a world of causation only by events, then we do not add any more control relevant to free will or moral responsibility. This is a powerful argument.

Applying similar reasoning when we think about OIC, we might wonder whether simply adding indeterminism will give us the robust ability to do otherwise that we need if we are to meet our moral obligations. If indeterminism simply makes it physically and metaphysically possible that a state of affairs obtains, but does not give us any particular control over it beyond what we have in a deterministic world, then it is not clear that it gives us all that we need to meet the ‘can’ condition operative in OIC. In other words, even if indeterminism is a necessary condition for being able to do something (that is, for meeting the ‘can’ condition) in the sense operative in OIC, it does not follow that it is sufficient. And given that, by Pereboom’s own lights, the ‘ought’ of moral obligation is action guiding, the ‘can’ in OIC is presumably a ‘can’ of control rather than simply a ‘can’ of bare possibility.

If this is right, then there is a good case to be made that the notion of control that is partially constitutive of free will in the basic desert sense is the very same notion of control that is operative in OIC. And if this is right in turn, then skepticism about free will goes hand in hand with skepticism about moral (and other) obligation. The challenge to the instantiation of moral obligation will not be limited to deterministic worlds, then, but to any worlds in which we do not meet the relevant conditions for control.5 Free will required for moral responsibility in the basic desert sense and the ability required for obligation need not be the same thing, for they might each require distinct additional necessary conditions. But there is reason to think that the control required for each is the same. And if adding indeterminism does not help achieve it in the one case, then, by parity of reasoning, it should not in the other.

Now the first thing to note here is that even if all of this were correct, the very same reply that Pereboom offers to the objection in its current form could still be offered to this more expansive version. If the forward-looking account of blame can incorporate the rejection of moral obligation and its replacement with the ‘ought’ of axiological recommendation, then it applies whether the world is deterministic or indeterministic. Still, the question of what the nature of the ‘can’ is that is operative in OIC is one of independent interest, and if free will skepticism goes hand-in-hand with obligation skepticism, it also means that the stakes might be even higher on the forward-looking account of blame that accommodates skepticism about moral obligation. Let us now turn to the implications of the (possible) rejection of moral obligation. I will focus most of my attention here on the implications for the treatment of criminals in the next section, but will also show how parallel questions can arise when it comes to the implications for personal relationships, and for the account of blame itself in section V.

IV. Implications for the Treatment of Criminals

In chapter 7, Pereboom assesses both what types of justification are available for the treatment of agents who engage in criminal behavior, and what types of treatment such justifications best support. It is immediately clear that no retributive justification of punishment will be available to the free will skeptic, but that is all to the good, on Pereboom’s view, in part for reasons quite independent of free will skepticism. This conclusion may naturally strike readers as an obvious and immediate consequence of free will skepticism, but Pereboom’s further conclusions each require further premises that are themselves in need of some elaboration. First, he rejects non-retributive justifications of punishment, including, among others, a utilitarian-based deterrence theory, and even a theory that appeals to the right to self-defense to justify punishment. Instead, he adopts the justification that functions in the latter theory, but argues forcefully that rather than supporting punishment, it serves only to justify incapacitation of criminals.

Before getting into the details of the argument, it will be helpful to distinguish between punishment and incapacitation. And it seems that for Pereboom, what is distinctive about punishment is that it is intended to inflict significant harm, such as death or severe physical or psychological suffering (p. 169). Although incapacitation, or preventative detention, might cause harm, that needn’t be even part of the intention. It is clear why a free will skeptic should reject retributivist justifications of punishment, but why reject retributivism’s long-standing rival, utilitarian-based deterrence theories? After all, such theories are notable precisely for not requiring that the one being punished deserve punishment; the justification is in terms of benefits to society that outweigh the costs. But as Pereboom points out, there are other general moral objections to such a theory, including the idea that the theory would seem to justify punishing the innocent (a practice that seems to be “more deeply wrong” than can be accounted for by offering utilitarian assurances that an institution that would do so would likely not be maximizing utility), as well as the idea that punishment on this theory would require using people without their consent to benefit others, something that is often “intuitively wrong” (p. 165). Now it is important to note here that Pereboom leaves the door open that this theory is correct, but makes the plausible methodological point that we would need more than mild confidence that it is, given the severe harms that punishment imposes.

Pereboom next considers a theory that appeals to the right of members of society to defend and protect themselves to justify punishment (see Farrell 1985; cf. Quinn 1985 and Kelly 2009). Pereboom argues persuasively that the theory does not justify punishment—understood as requiring the intentional imposition of harm—so much as it supports the detention and incapacitation of criminals, which does not require intended harm. On Pereboom’s positive view, rather than punishment, the right response to criminal wrongdoing (and perhaps to excellent evidence of future wrongdoing) is best understood on the analogy to quarantine of those with communicable diseases (p. 169, and see Schoeman 1979). But like the rights-based theory of punishment, it, too, is based on “our right to protect ourselves and to secure our safety” (p. 169). For even if we do not aim at harm, harm is often a foreseen side-effect of incapacitation, and needs to be justified when it is undeserved. It is the right to self-defense that provides this justification.

Now we can begin to see one way that this view and the arguments for it are connected to the earlier discussion of moral obligation. Since the view invokes rights, the question now arises whether it is possible to have rights in the absence of duties—or, in other words, in the absence of obligations on the part of others not to interfere with one’s exercise of said rights. While we can understand what it is to have duties toward individuals who may lack rights (e.g., toward certain non-human animals), the very idea of a right would seem to include the idea of correlative duties. In that case, it might be doubted whether the earlier acceptance of the possibility that no one has moral obligations is consistent with the availability of rights as a potential justification here. Thus, there is an apparent tension between three claims:

1. 1. It is possible that no one has any moral obligations, or moral duties, in our world.
2. 2. We have rights, including to self-defense and liberty.
3. 3. Rights require correlative duties.

Now (1), (2), and (3) do not together constitute a contradiction. (1) is only a possibility claim. Suppose, though, that determinism is true, and so, by Pereboom’s reasoning, no one has any moral obligations. Suppose then, that

(1’) No one has any moral obligations in our world.

We still do not have a contradiction. It would only follow from (1’) and (3) that (2) is false, and, thus, we lack rights if we understand (3) in one of a set of particular ways. For example, we could read (3) as claiming that A’s having rights requires that there actually exist other individuals who have correlative duties, in which case, we would appear to be able to generate a contradiction. But this way of reading (3) seems dubious. If A is the only person in the world at the moment, it is doubtful that she has lost her rights and would regain them only when other people come into existence. Still, it seems that having rights supports the truth of certain counterfactuals. Were there other agents capable of having duties, they would have duties not to interfere in A’s exercise of her right to life, for example. If we read (3) in this way, then it becomes very important whether the reason for (1’)—if it is true—is empirical and contingent, or whether there is some incoherence in the very idea of the control required to have such obligations. If it is merely a contingent fact like there being no other rational agents in the world, then this would not impugn the existence of rights. But if there is some conceptual incoherence in the idea of a moral duty, then it would seem that the same would be true for the idea of a right. The situation we are to imagine here, namely, that (1’) is true because determinism is true, poses an interesting question. Could we have rights in a world in which it was, in at least some sense, not possible to have obligations? If it is logically possible to have duties, but not physically or metaphysically, could we have rights? Is this case relevantly like the first case in which there happens to be only one person in existence or like the second case in which the concept is incoherent? To claim that it is more like the first case would require seeing the obstacle to the ability to do otherwise as empirical. That in turn would seem to depend on seeing the obstacle as simply determinism (contrary to the way I was urging us to see it in the previous section), or as an otherwise merely contingent failure to be agent causes (thereby committing to the ultimate conceptual coherence of the relevant kind of control for the ability to do otherwise, something that Pereboom has left as an open question).

Relatedly, one might also wonder at this point what the normative force is of the incapacitation theory itself. It cannot be (or at least if determinism is true, it cannot be) that we ought to—in the sense of have a moral obligation to, or of specific action demand—incapacitate rather than punish, say. And if it turns out that we are only able to invoke the ‘ought’ of ‘it would be better if,’ then it is not clear that any non-consequentialist considerations can come into play for us. Does this mean that Pereboom adopts a consequentialist position after all?

Here, too, there is an apparent tension, but one that I believe is only apparent in the end. On the one hand, there are appeals to what seem like straightforwardly non-consequentialist ideas. It seems to matter, for example, whether harm is intended (as it is in punishment) or merely foreseen (as it is in incapacitation). Appeal to a morally significant distinction between intended and foreseen harm is a classic non-consequentialist idea. The appeal to rights is itself traditionally a non-consequentialist idea, as is an appeal to fairness (see p. 171, for example). Finally, recall Pereboom’s claim that it seems more deeply wrong to punish the innocent than can be accommodated on a utilitarian view. On the other hand, when it comes to the explicit adoption of a moral theory, Pereboom expresses his “…predilection to endorse an axiological moral theory which includes better consequences as valuable, where morally fundamental rights being honored and not violated count among the good consequences” (p. 172). This sounds like a consequentialist view—albeit without the common appeal to our moral obligation to maximize valuable consequences.

The apparent tension can be resolved in at least two ways, however. First, recall that Pereboom only claims that punishing the innocent seems to be more deeply wrong than can be accommodated even on sophisticated utilitarian views. Even more importantly, Pereboom adopts a sophisticated rights consequentialist theory, according to which rights violations themselves count as bad consequences. Thus, we can preserve much of our talk of rights and fairness and even an intend/foresee distinction by recognizing these as figuring in various ways in the value of the consequences of our actions.

The tension is thereby resolved, but at the same time, it is worth bringing out that the resulting theory is revisionary for our practices on two very significant counts. First, even insofar as it is a consequentialist theory, it is radically different from most in not having a role for the moral ‘ought’ of action demand. Second, one is always in danger of ending up in a situation in which “it would be better” to violate a right to prevent even more violations (or even to prevent an equal number of violations plus one slight harm), and this is at least in a way quite radical. To borrow an analogy from Philippa Foot, moral prescriptions would take a form more like “see to it that there is no shouting” which could in fact be accomplished by shouting, than like “don’t shout” (1985/2003, p. 99).6 Applied to the case at hand, we should say “see to it that there are fewer rights violations,” not “do not violate rights.” (Of course, in a particular case we could recommend satisfying this general axiological recommendation by not violating rights ourselves. But the specific recommendation would get its force from the general idea that it would be better if there were fewer rights violations). This is not by itself sufficient reason to reject the view, and a full evaluation of this particular kind of consequentialism would require at least a book of its own. But it allows us to see, as Foot and others have argued, that such a view requires significant revision of our view of morality itself. In addition, the purely axiological approach will read the “see to it that” only as a recommendation to do what is valuable. Thus, skepticism about the ‘ought’ of specific action demand yields what seems to be a significantly revisionary approach to our moral practices, and we see it play out here in the justification of the treatment of criminals.7

Finally, it is worth noting that the approach to morality would seem to rule out another non-retributivist account of the treatment of criminals, one that rests on the idea that wrongdoers incur obligations of certain kinds, including the obligation to pay compensation, but also perhaps a further obligation to make it right in some way or rectify the wrong in the moral case (see Nelkin 2013), or a further obligation to contribute to the protection of society in the legal case (see Tadros 2011).8 If the obligations incurred are of a kind that would be served by incapacitation, then incapacitation can be justified by appeal to an enforceable duty on the part of the offender on such a view. I find this general approach appealing in certain ways, but it would seem to be ruled out—at least in a deterministic world—by Pereboom’s approach.

In sum, Pereboom makes a persuasive argument that one can justify a reasonable treatment of criminals without appeal to free will required for basic desert, and shows that regardless of free will skepticism, there is good reason to accept his approach. But perhaps the even more radical idea here is that one can do it without appeal to obligation, without the ‘ought’ of specific action demand and the moral framework that seems to depend on it.

V. Implications for Blame and Relationships

As I tried to show in the previous section, when it comes to the treatment of criminals, we could accept Pereboom’s arguments that we can avoid appealing to free will and desert for justification, but still have doubts about whether we can avoid appealing to obligation. I think something similar is true when it comes to both blame and to valuable personal relationships. I will simply sketch how I think this works in each case, and leave elaboration for another occasion.9

First, consider the forward-looking account of blame. As Pereboom notes (following McKenna), the process of blame might itself be harmful to the wrongdoer and include, among other things, psychological pain (p. 134). According to Pereboom, “inflicting this harm is justified first of all by the right of the defamed to protect themselves and to be protected from this type of aggression and its consequences.” While the infliction of harm is not an intrinsic good, and not the point of the blame, on Pereboom’s view, it still needs justification, and that comes (in part) in the form of the rights of the victims of wrongdoing. Thus, in the same way that questions arise about a role for rights given acceptance of the possibility that no one has moral obligations, a similar question arises here.

Next, consider personal relationships. Here Pereboom does a masterful job building a case that we can have the most valuable of human relationships without a susceptibility to the reactive attitudes like resentment and indignation, or at least we can do so while taking significant steps towards eliminating and disavowing such attitudes. There is much that is interesting to consider about Pereboom’s insights and arguments here. But in keeping with the aim of bringing out implications of the elaborated forward-looking account of blame, I want to add another dimension to the mix here and that is the element of obligation. Arguably, relationships like friendship are defined in part by the special obligations friends have to each other. Two people might like each other, enjoy each other’s company, and wish each other well but still not be friends. Intuitively, it seems to me, at least part of what could be preventing this relationship from being a friendship is a lack of special obligations to each other. To see this, consider a situation in which one of the pair is in need. Now suppose that one did not have an obligation to help the other that was in any way different from an obligation to help a stranger in a similar situation. I think we would then have good reason to question whether the two were friends. Of course, the nature of friendship is also a topic worthy of more space than I can give it here, but my aim is simply to bring out the fact that skepticism about obligation, or “ought” in the specific action demand sense, challenges more than certain assumptions about morality; it threatens to undermine the existence of friendship as we think about it, as well.

Now perhaps not coincidentally, in facing this sort of objection, I think Pereboom finds himself in company with consequentialists. For while most do not deny the existence of obligations per se, they do tend to deny the existence of special obligations. (See Diane Jeske 2014 for a very helpful overview.). So if friendship requires such a thing, then consequentialism would have the very unwelcome consequence that there is no such thing as friendship. One option at this point would be to join consequentialists (or others) who deny that friendship requires special obligations after all and attempt to offer an alternative conception of friendship.10 As before, I won’t attempt a full evaluation of the prospects of this strategy here; my aim is simply to bring out the potential challenge arising from a sort of skepticism about obligation. The resources of the “ought” of axiological recommendation can be brought in when thinking about friendship as well as morality (and, indeed, see Pereboom 2014b). But it seems to me that a challenge remains to account for friendship without the existence of special obligations.

VI. Conclusion

In each of the cases—criminal treatment, blame, and personal relationships—I find myself tempted to agree that it is not a lack of desert per se that is most threatening to the things we value most and that seem so central to our way of life. Rather, it is a lack of something intimately related to, but—on my view, not Pereboom’s—even more fundamental than desert that is most threatening, and that is the lack of obligation. But the remarkable thing is that Pereboom has shown that there are resources available to try to meet even this threat.

Living Without Free Will was a tremendous achievement in part for showing how much we could retain even if we are not free in a basic desert sense. In Free Will, Agency and Meaning in Life, Pereboom has only deepened his reasoning in favor of this optimistic view. And in addition to his arguments for free will skepticism itself, he has extended his view so that while he may have even less to work with (in being even more committed here to the need to offer an alternative picture to that which appeals to moral obligation), he still manages to offer a deep and subtle case for a hopeful view of human life. While I have given reasons why I don’t share quite the same level of confidence that we can do so much with so little, there should be no doubt that Pereboom’s work brings out in completely new ways a rich sense of what the possibilities are.

Acknowledgment

Many thanks to Gregg Caruso, Derk Pereboom, and Sam Rickless.

References

• • Arneson, Richard (2003). “Consequentialism vs. Special-Ties Partiality,” The Monist 86: 382-401.
• • Caruso, Gregg. “Précis of Free Will, Agency, and Meaning in Life,” Science, Religion and Culture 3: 178-201.
• • Fischer, John Martin (2003). “’Ought-Implies-Can’, Causal Determinism, and Moral Responsibility,” Analysis 63: 244-50.
• • Foot, Philippa (1985/2003). “Morality, Action, and Outcome,” Objectivity and Value: Essays in Memory of John Mackie (London: Routledge & Kegan Paul). Reprinted in Philippa Foot, Moral Dilemmas and Other Topics in Moral Philosophy (New York: Oxford University Press): 88-104.
• • Haji, Ishtiyaque (1999). “Moral Anchors and Control,” Canadian Journal of Philosophy 29: 179-203.
• • Jeske, Diane (2014). “Special Obligations”, The Stanford Encyclopedia of Philosophy (Spring 2014 Edition), Edward N. Zalta (ed.), URL = <http://plato.stanford.edu/archives/spr2014/entries/special-obligations/>.
• • Levy, Neil (2012). “Skepticism and Sanction: The Benefits of Rejecting Moral Responsibility,” Law and Philosophy 31: 477-93.
• • McKenna, Michael (2012). Conversation and Responsibility (New York: Oxford University Press).
• • Nelkin, Dana Kay (forthcoming). “Friendship, Freedom, and Special Obligations,” in Agency and Moral Responsibility, eds. Andrei Buckareff, Carlos Moya, and Sergi Rosell (Palgrave-MacMillan).
• • Nelkin, Dana Kay (2013). “Freedom and Forgiveness,” in Free Will and Moral Responsibility, Ishtiyaque Haji and Justin Caouette (Newcastle upon Tyne: Cambridge Scholars Press).
• • Pereboom, Derk (2014a). Free Will, Agency, and Meaning in Life (Oxford: Oxford University Press).
• • Pereboom, Derk (2014b). “The Dialectic of Selfhood and the Significance of Free Will,” in Libertarian Free Will: Essays for Robert Kane, ed. David Palmer (New York: Oxford University Press).
• • Pereboom, Derk (2001). Living Without Free Will (Cambridge: Cambridge University Press).
• • Sen, Amartya (1981). “Rights and Agency,” Philosophy and Public Affairs 11: 3-39.
• • Smilansky, Saul (2011). “Hard Determinism and Punishment: A Practical Reductio,” Law and Philosophy 30: 353-67.
• • Tadros, Victor (2011). The Ends of Harm: The Moral Foundations of Criminal Law (Oxford: Oxford University Press).

Endnotes

1 One qualification is needed here: the thesis Pereboom accepts in not strictly the denial of the original Principle of Alternate Possibilities. It is that one can be responsible for what one does without having robust alternatives, that is, without having alternatives that have a role per se in explaining why one is responsible. According to Pereboom, responsibility—and blameworthiness—might very well require alternatives because they are entailed by indeterminism. But one is not responsible even partly in virtue of having alternatives (p. 28). Thus, Pereboom rejects what we might call the “Principle of Robust Alternate Possibilities” according to which one is blameworthy for what one has done only if one had a robust alternative.

2 See McKenna (2012). McKenna argues that his conversational model is consistent with a justification based on desert and with one that is not so based.

3 Again, technically, Pereboom rejects a close variant of PAP-B. (See note 1.)

4 These accounts include those of C.D. Broad, Ruth Barcan Marcus, Lloyd Humberstone, Gilbert Harman, Kate Manne, and James Hobbes.

5 Interestingly, at one point in Pereboom (2001)considers something very like this expanded objection, writing that even if determinism were false but our “choices and actions were partly or truly random events”, then we still might not have the relevant ability to do otherwise required by the “ought” of moral obligation (p. 143). See also Ishtyiyaque Haji (1999) for an argument that moral obligation might be threatened not only in a deterministic world, but also in three sets of indeterministic conditions that have been proposed as sufficient for moral responsibility.

6 Foot is, in part, responding to the “goal rights system” put forward by Amartya Sen (1981).

7 Pereboom uses the view to good advantage in responding to an objection of Saul Smilansky’s that we would owe criminals not only a lack of gratuitous harm, but also a lavish compensation package. Smilansky (2011) argues that one of the central motivations for skepticism about free will is a concern for fairness, and that the very idea of fairness is undermined in a system in which we justifiably cause significant harm through incapacitation without such impossible-to-attain compensation. In response, Pereboom writes approvingly of Neil Levy’s (2012) appeal to a kind of consequentialism that incorporates such traditionally deontological notions of rights and fairness into the consequences themselves.

8 Perhaps, though, this approach, while non-retributive, still depends on the wrongdoing being culpable, in which case Pereboom could reject this approach on these grounds alone, as long as the culpability here entails basic desert. This is an interesting question that I hope to explore on a future occasion.

9 I discuss Pereboom’s work on relationships in more detail in Nelkin (forthcoming).

10 See Richard Arneson (2003) for a subtle working out of such a view.